Research map / Detection and segmentation / CNN features
Learning Monocular Depth in Dynamic Scenes via Instance-Aware Projection Consistency
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add instance-wise geometric consistency to monocular depth and motion learning (ablation) | Insta-DM · KITTI Eigen split | AbsRel (all): 0.156 -> 0.137 | Helped |
| Use forward instead of inverse warping for object regions (ablation) | Insta-DM · KITTI Eigen split | AbsRel (all): 0.146 -> 0.124 | Helped |
| Use forward instead of inverse warping for object regions (ablation) | Insta-DM · KITTI Eigen split objects | AbsRel: 0.362 -> 0.178 | Helped |
| Add geometric consistency when object regions use inverse warping (ablation) | Insta-DM · KITTI Eigen split | AbsRel (all): 0.137 -> 0.146 | Hurt |
| Use inverse warping for object regions (ablation) | Insta-DM · KITTI Eigen split object regions | object motion convergence | Hurt |
| Train with instance-aware projection and consistency | Insta-DM with DispResNet · KITTI Eigen test | AbsRel: 0.137 -> 0.124 | Helped |
| Train with instance-aware projection and consistency | Insta-DM with ResNet18 · KITTI Eigen test (semantic knowledge) | AbsRel: 0.113 -> 0.112 | Helped |
| Train with instance-aware projection and consistency | Insta-DM · Cityscapes test | AbsRel: 0.119 -> 0.111 | Helped |
| Train with instance-aware projection and consistency | Insta-DM · KITTI 2015 scene flow training set | D1 fg.: 34.4 -> 30.4 | Helped |
| Use instance masks while training Ego-PoseNet (ablation) | Insta-DM · KITTI visual odometry Seq. 09 | ATE: $0.012\pm0.008$ -> $0.010\pm0.013$ | Mixed |
| Use instance masks while training Ego-PoseNet (ablation) | Insta-DM · KITTI visual odometry Seq. 10 | ATE: $0.011\pm0.010$ -> $0.011\pm0.008$ | No clear effect |
Closest papers by meaning
- Mining Supervision for Dynamic Regions in Self-Supervised Monocular Depth Estimation (CVPR 2024 · similarity 0.91)
- From-Ground-To-Objects: Coarse-to-Fine Self-supervised Monocular Depth Estimation of Dynamic Objects with Ground Contact Prior (CVPR 2024 · similarity 0.89)
- Seurat: From Moving Points to Depth (CVPR 2025 · similarity 0.88)
- MonoInstance: Enhancing Monocular Priors via Multi-view Instance Alignment for Neural Rendering and Reconstruction (CVPR 2025 · similarity 0.88)
- Zero-Shot Monocular Scene Flow Estimation in the Wild (CVPR 2025 · similarity 0.88)
- Depth2Pose: A Pose-Based Benchmark for Monocular Depth Estimation without Ground-Truth Depth (NeurIPS 2026 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.